Interpretation of the latest Second Life service quality metrics
I made some interpretations of the latest Second Life service quality metrics.
This graph shows us the planned & unplanned outages.
I would interpret from this, knowing that there was lots of Denial of service attacks on SL grid after major bad news (Gambling, VAT, Ginko), that during July and August, there was a lot of attacks on the grid, while overall grid has been becoming way more stable since March.
After April / May, either less grid updates has been done OR they have been more successful than in the past, requiring less of downtime. Also, July had higher downtime probably because of “Message Liberation”, meaning releasing the groundwork for HET grid, at least, that i would believe be the case for higher planned downtime.
This is basically how often viewer crashes!
I’m not going to cross reference this to user spent hours, just quickly state that it seems while some has been fixed, new problems are regularly introduced.
However, how come there isn’t a spike for July & August? I would have guessed then would have been a lot more due to so much unplanned outage causing booting users.
This is REALLY bad, over 20% of sessions ending into a crash! Ouch! That got to be felt on the user base too!
How fast does your SL run?
This figure is a major OUCH! Really, that slow? I mean come on, human eye perceives 24 frames per second! So everyone feels ´lagged´ all the time? Apparently so.
For example, in first person shooters, you need even higher FPS for smooth experience, and online, you feel the difference between 60 and 40FPS.
Some major optimizations got done early in the year, but then new stuff has been introduced for a lot of the advantage disappear! While if no changes are done, theoretically average FPS should slowly increase as computers get faster.
Because average users so rarely upgrade and short time span, that increase of ~1 to 3% in speed doesn’t count much. (In actuality, computers get faster, well, way faster, but computers get upgraded not so often).
That is a major headache. Image the 25% of people running at under 8 oO; Wow, that is majorly slow, and must be majorly annoying. Why we do not see more attempts to increase FPS?
This is actually quite good news, knowing that Second Life is very heavy, this figure is so good! Most of the time hiccups aren’t server side.
But then again, most of second life is always empty soooo … This does not say much, how about statistics for sims while atleast 5 or 10 avatars present? That would be so much more useful graph.
Empty sims shouldn’t lag, right? So, the tons and tons of sims empty might make this figure unusable.
That concludes my thoughts of these graphs … Of course, one could make deeper analysis and really get into it, but the idea of this post wasn’t to overly analyse these graphs 
So, what does these graphs bring into your mind?